»  Home »  Bookstore »  Title index »  Categorical, count, and censored outcomes »  Multivariable Model-Building: A Pragmatic Approach to Regression Analysis Based on Fractional Polynomials for Modelling Continuous Variables

## Multivariable Model-Building: A Pragmatic Approach to Regression Analysis Based on Fractional Polynomials for Modelling Continuous Variables

Click to enlarge
See the back cover

What are VitalSource eBooks?
Your access code will be emailed upon purchase.

eBook not available for this title

eBook not available for this title

 Authors: Patrick Royston and Willi Sauerbrei Publisher: Wiley Copyright: 2008 ISBN-13: 978-0-470-02842-1 Pages: 322; hardcover
 Authors: Patrick Royston and Willi Sauerbrei Publisher: Wiley Copyright: 2008 ISBN-13: Pages: 322; eBook Price: $0.00  Authors: Patrick Royston and Willi Sauerbrei Publisher: Wiley Copyright: 2008 ISBN-13: Pages: 322; Kindle Price:$

Review of this book from the Stata Journal

 Supplements: Errata, data, and other materials

### Comment from the Stata technical group

Selecting the appropriate model from among a large class of candidate models is a difficult process: one must balance the (sometimes contradictory) goals of model interpretability, parsimony, good prediction properties, robustness to minor variations in the data, and applicability to other data. This text presents a well-rounded, practical approach to model selection, with its bulk devoted to general variable selection through the use of stepwise procedures (or otherwise) and the selection of functional forms for continuous variables. Regarding the selection of functional forms, the authors pay much attention to fractional polynomials and splines, drawing on their vast research in these areas. In particular, those looking for a tutorial on the use of fractional polynomials will find this text very useful. The methods prescribed can be applied widely, yet the examples used are primarily from the health sciences, with the typically used models being logistic regression, Cox regression, and generalized linear models.